Bon il y a bien deux trois trucs qu'on sait prouver quand même, mais oui 😅

Le crois que t'as mal orthographié "Mistral" 😛

Pour les formules arithmetiques il semblerait qu'il y aie une version intuitionniste de la hiérarchie arithmétique: www.cambridge.org/core/books/a...

C'est aussi mentionné dans la deuxième moitié de cette section: plato.stanford.edu/archives/sum...

What's the other 40%?

Oui d'ailleurs je me rappelle que l'établissement dans lequel j'ai passé mon bac était un lycée privé sous contrat. Dans la salle où j'ai passé mes épreuves, il y avait une croix chrétienne. J'avoue que sur le coup ça fait bizarre quand on a fait que du publique 😅

And unless they need it as an intermediary step in one of their proofs they wouldn't bother proving it for the sake of proving it.

At least this is how I understand their attitude 😊

6/6

I believe they have a similar attitude towards the existance of 2^n. They believe it's probably true (because of standard proofs, because they have a good idea of what algorithm they woud need run to check it), though they do not considered it proven without a full calculation. 5/6

but for which if they showed as a small intermediary claim in one of your papers, you would not feel comfortable claiming them true without actually providing a proof.

Not all such statements are necesseraly worth investigating on their own. It might even be a boring statement. 4/6

their existence.

Now, there certainly are statements you think are probably true but that you have not proven. Maybe because the statement is not that interesting, or because you just have not taken the time to do so.

These could be the kind of statements you are almost sure holds, 3/6

But it does not mean they beleive the large numbers don't exist. Just that before seeing the whole computation they would not consider we proved it. As such, they might very well think that those numbers *probably* exist, but just that the common proofs are not sufficient to mathematically claim 2/6

Hmm, to me the ultrafinitists are like paranoid mathematicians. They need the absolute certainty that something exists to claim it does. Not only do they need a method to compute the object, but also be able to run it an get the output of the computation. 1/6

Non mais elle assume donc ça va!

The above skeet is offered to you by the ministry of vacuous truth.

Surtout que ces cas extrêmes existent: github.com/microsoft/Bi...

Est ce qu'il y a eu qqch de formalisé par l'IA qui n'a jamais été formalisé avant ?

Pour moi l'approche la plus prometteuse (mais pas vraiment applicable aux navigateurs) c'est que le régulateur impose l'interopérabilité adversariale: en gros l'acteur dominant est forcé d'ouvrir son produit aux autres. www.eff.org/deeplinks/20...